Description of problem:

I noticed that clone()d thread sleeps forever on exit.
The clone()d thread's state becomes zombie, but it is not notified to
the process calling waitpid().

Not applying the three patches listed below solves this problem.
I think this problem is introduced by one of these patches.

  # Roland's utrace ptrace replacement.
  ApplyPatch linux-2.6-ptrace-cleanup.patch
  ApplyPatch linux-2.6-tracehook.patch
  ApplyPatch linux-2.6-utrace.patch

Version-Release number of selected component (if applicable):

Fedora 8's 2.6.26.3-14.fc8 and Fedora 9's 2.6.26.3-29.fc9 are affected.

How reproducible:
100%

Steps to Reproduce:
1. Compile the following program and run.

----------
#define _GNU_SOURCE
#include <sched.h>
#include <sys/types.h>
#include <sys/wait.h>
#include <stdio.h>
#include <errno.h>
#include <stdlib.h>

static int child(void *arg) {
	return 0;
}

int main(int argc, char *argv[]) {
	int error;
	char *stack = malloc(8192);
	printf("Calling clone() ");
	{
		const pid_t pid = clone(child, stack + (8192 / 2), CLONE_NEWNS, NULL);
		printf("(pid=%d)\n", pid);
		fflush(stdout);
		while (waitpid(pid, &error, __WALL) == EOF && errno == EINTR);
	}
	printf("waitpid() returned %d\n", WIFEXITED(error) ? WEXITSTATUS(error) : -1);
	return 0;
}
----------
2.
3.
  
Actual results:
It sleeps forever until terminated by Ctrl-C.

Expected results:
It sould return immediately.

Additional info: